【摘要】 提出适用于柑橘砧木育苗的枳壳种子离散元模型构建方法，并系统标定相关离散元参数。针对枳壳种子外形不规则的特点，按形态特征将其划分为椭球形、扁圆形和纤瘦形三类。随机选取2 000粒种子样本，测定其三轴尺寸及形状组成比例；通过堆积角试验与轴向压缩试验获取种子的堆积角与剪切模量。以典型种子外轮廓三维面片模型为几何约束，采用球状颗粒填充算法，构建三类表征真实立体形态的种子颗粒模型。基于“Hertz-Mindlin with JKR”接触模型，建立模拟种子群体堆积与筛分过程的离散元模型。以种子间及种子—钢板接触参数(如恢复系数、摩擦系数)和JKR表面能为因素，圆筒提升法实测堆积角为响应目标，通过Plackett—Burman试验、最陡爬坡试验和Box—Behnken试验，构建堆积角与关键参数的二阶回归模型。以试验堆积角为目标值，优化求解最优参数组合，并进行筛分验证试验。PB与Box—Behnken试验表明，枳壳种子—不锈钢碰撞恢复系数、滚动摩擦系数及JKR表面能是影响堆积角的主因。优化得参数最优值分别为0.39、0.28和0.90 J/m~2。仿真堆积角与物理试验相对误差为2.96%,堆积形态高度相似；仿真透筛率与实际试验相对误差为5.49%,验证参数准确性。所建模型及参数能准确模拟种子的堆积、流动与分离行为，为分析种子—播种装置相互作用及设计高效精准播种器提供理论依据。更多还原

【Abstract】 This study aims to propose a method for constructing a discrete element model(DEM) of Poncirus trifoliata seeds, commonly used in citrus rootstock cultivation, and systematically calibrate the relevant DEM parameters. To address the irregular shape of the seeds, they were categorized into three morphological types: ellipsoidal, oblate, and slender. A random sample of 2 000 seeds was selected to measure their triaxial dimensions and the proportional distribution of the three shapes. The angle of repose and shear modulus of the seeds were determined through physical experiments, including angle of repose tests and axial compression tests. Based on the 3D surface models of typical seed outlines, spherical particle filling algorithms were employed to construct three types of seed particle models representing realistic morphological features. Using the “Hertz-Mindlin with JKR” contact model, a DEM was established to simulate the stacking and screening processes of seed assemblies. Key parameters, including seed-seed and seed-steel contact parameters(e.g., restitution coefficient, friction coefficient) and JKR surface energy, were considered as influencing factors, with the angle of repose measured by the cylinder-lifting method as the response variable. A series of experiments—Plackett—Burman(PB) design, steepest ascent, and Box—Behnken design—were conducted to develop a second-order regression model relating the angle of repose to critical DEM parameters. The optimal parameter set was determined by optimizing the model against the experimentally measured angle of repose, and a screening process validation test was performed. Both PB and Box—Behnken tests indicated that the restitution coefficient between seeds and stainless steel, the rolling friction coefficient, and the JKR surface energy were the primary factors affecting the angle of repose. The optimized values for these parameters were 0.39, 0.28, and 0.90 J/m~2, respectively. Simulations using these parameters yielded a stacking angle with a relative error of 2.96% compared to physical tests, and the stacking morphology showed high similarity. The relative error between the simulated screening efficiency and the actual test results was 5.49%, validating the accuracy of the calibrated parameters. The established DEM and calibrated parameters accurately simulate the stacking, flow, and separation behaviors of Poncirus trifoliata seeds. This study provides a theoretical foundation for analyzing the complex interactions between seeds and seeding devices and supports the design of efficient and precise citrus seeding equipment.更多还原